AI regulation in Australia: the approach in brief
Adopted 2026-06-16 · ≈ 2 min read · 
edited by Dirk Baaijen
Australia does not regulate AI through a single horizontal law. The system rests on voluntary ethics principles, a voluntary safety standard and a proposal for mandatory guardrails for high-risk AI, alongside existing sectoral law.
Short answer: Australia has no horizontal AI law like the EU. Its approach rests on voluntary instruments — the AI Ethics Principles and the Voluntary AI Safety Standard — complemented by a government proposal for mandatory guardrails for AI in high-risk settings. Meanwhile, existing law on areas such as privacy, consumer protection and anti-discrimination continues to apply in full.
Voluntary principles as the foundation
At the core of Australian policy are the Artificial Intelligence Ethics Principles, published in 2019 by the Department of Industry, Science and Resources. These are eight voluntary principles — including human, social and environmental wellbeing, fairness, privacy protection, transparency, contestability and accountability — intended to help organisations design and deploy AI responsibly. The principles are not legally binding and have no regulator with enforcement powers; they serve as guidance.
A voluntary safety standard
To make the principles operational, the National AI Centre issued the Voluntary AI Safety Standard. The standard translates the ethics principles into concrete guardrails for organisations that develop or use AI, addressing risk management, transparency, human oversight and accountability across the supply chain. This standard is also voluntary, but is designed to align with any future mandatory regime.
Proposal for mandatory guardrails for high-risk AI
Through the Department of Industry, the Australian government has consulted on mandatory guardrails for AI in high-risk settings. The proposal is risk-based: heavier obligations — such as testing, transparency and accountability requirements — would apply to applications with a higher risk of harm, while low-risk applications would remain under the lighter voluntary track. This mirrors the international shift toward risk-based regulation, but the precise scope and legislative instrument are still being developed; consult the official consultation page for the current state of play.
Existing law still applies
Even without a horizontal AI law, AI in Australia does not operate in a legal vacuum. Existing rules on privacy, consumer protection, competition, online safety and discrimination apply to AI systems. The proposed mandatory guardrails are intended to fill the gaps these sectoral frameworks leave, not to replace them.
Read more: International AI governance. Take the scan.
Sources
- https://www.industry.gov.au/publications/australias-artificial-intelligence-ethics-principles
Australian government page setting out the eight voluntary AI Ethics Principles (published 2019). - https://consult.industry.gov.au/ai-mandatory-guardrails
Consultation page of the Department of Industry on mandatory guardrails for AI in high-risk settings. - https://www.industry.gov.au/publications/voluntary-ai-safety-standard
The Voluntary AI Safety Standard from the National AI Centre, aimed at responsible AI use by organisations.
Read next
The Algorithm Register for Dutch public authorities
The Dutch Algorithm Register (algoritmes.overheid.nl) is the central public platform where government organisations voluntarily publish information about the algorithms they use; registration in the Dutch register is not legally mandatory, but public authorities deploying high-risk AI systems…
MDR and the AI Act: regulatory overlap for medical software
Medical software with AI components falls simultaneously under the MDR and the AI Act; the high-risk classification via the MDR route (Article 6(1) AI Act) does not apply until 2 August 2027.
Government Algorithm Transparency
The EU AI Act (2024/1689) requires public authorities deploying high-risk AI systems to register them in an EU database, document their operation transparently, and notify affected individuals; the Netherlands leads with a voluntary Algorithm Register expected to become legally mandatory.